Early Detection of Rare Inherited Retinal Dystrophies and Cardiac Amyloidosis enhanced by Artificial Intelligence: the impact on the patient's pathway in Campania Region

Project Objectives

In Europe, rare diseases are defined as conditions affecting less than 5 individuals in 10.000. Their diagnoses are challenging, requiring ultra-specialized expertise and devices, which are costly and available only in ad hoc referral centres. Most patients affected by rare disease are correctly diagnosed only after years and many visits, since general practitioners, paediatricians or specialistic physicians may lack dedicated knowledge and equipment for rare diseases investigation. Cardiac Amyloidosis (CA) and Inherited Retinal Dystrophies (IRDs) represent emblematic examples of rare diseases with a growing interest due to the great effort for development of novel therapies, already in the market, or currently in phase I, II, III clinical trials. Artificial intelligence (AI) is increasingly applied to medicine, and particularly to cardiology and ophthalmology, for improving effectiveness and cost-effectiveness of diagnoses, while reducing errors and costs. Pioneering studies applied AI to CA and IRD revealing the enormous potential of these approaches, in all the steps of the patient’s clinical pathway (i.e., PDTA: percorso diagnostico terapeutico assistenziale), reducing diagnostic delay, and ensuring appropriate therapies to all patients.

Thus, specific aims of this project will be:

• to develop an AI-based application for screening of CA;

• to pave the way to the development of an AI-integrated handheld fundus camera to support screening of IRDs in non-specialised clinics;

• to assess the impact of AI applications for enhanced referral and computer assisted diagnoses (CADs) of rare disease (cost-effectiveness and Health Technology Assessment of AI for CA referral and IRD CADs).

The project consortium includes four Units:

• Unit 1: A.O.R.N. Dei Colli - Ospedale Monaldi, Unità Di Malattie Rare e genetiche Cardiovascolari
• Unit 2: AOU Università degli studi della Campania Luigi Vanvitelli
• Unit 3: Università Campus Bio-Medico di Roma
• Unit 4: AOU Federico II

Unit 1 (Monaldi) will run a clinical investigation with an observational cohort study aiming to deploy a high-quality dataset useful to train and test AI models for CA referral.

Unit 2 (Vanvitelli) will define and conduct a pilot clinical study with a case control design to deploy a dataset useful to train and test the AI models.

Unit 3 (UCBM) will be responsible for the AI modelling.

Unit 4 (Federico II) will be responsible for patient’s genotyping.

Our hypothesis is that AI can enhance rare disease referral and CAD, making diagnosis more affordable, effective and cost-effective, while reducing errors. We will focus on CA and IRD, which may lead to chronic conditions (e.g., congestive heart failure or blindness), which are highly detrimental for patients¿ (and their families) quality of life and extremely costly for the National Health System (NHS) and society, if not prevented.